How does predictive analytics fit in a pipeline operator's maintenance strategy?
As midstream companies progress through their digital transformation journey, opportunity is being uncovered on how to better use their data to drive maintenance related decisions. In most cases these decisions have historically been driven by time-based maintenance schedules or condition based maintenance. By leveraging tools that can turn data into insight, the opportunity to move from a more reactive maintenance strategy to a proactive strategy becomes a reality.
The industrial Internet of Things and Industrie 4.0 are concepts that encompass various toolsets with the ability to leverage data to drive operational safety and profitability improvements along with improved asset reliability. And these benefits are resonating with pipeline operators. Predictive analytics enables operators to mine data for equipment behavior patterns that may indicate a problem the operator otherwise wouldn’t have seen. For example, predictive asset analytics can identify when rotational assets like pumps and compressors may be trending towards failure.
Rotating equipment at compressor and pump stations including gas turbines and centrifugal compressors and pumps require regular maintenance. This could include routine maintenance, inspections or a complete overhaul depending on the maintenance schedule. A study by ARC Advisory shows that only 18% of asset failure is age related. This is really where a more proactive (including predictive) approach to maintenance becomes essential.
Watch our recent webinar that introduces predictive analytics for pipeline operators, how it works and the types of failures that can be detected.
Being able to collect, monitor and analyze data such as bearing temperature, vibration and reduction in hydraulic performance enables operators to identify early signs of a possible failure giving a pipeline operator the opportunity to schedule planned maintenance instead of reactively needing to shut down all or a part of a pipeline impacting their operations both up and downstream.
What steps are operators taking to see if a predictive approach works for them?
One approach that many operators take with regards to predictive maintenance is what could be called “dipping your toe in the water”; A pilot (sometimes called an offline trial) or equivalent is used to validate that there is going to be a benefit from introducing this into an operator’s maintenance program.
A pilot typically takes the form of providing historical data for a few pumps or compressors that can be used to train the predictive model. This is then used against some historical data for the equipment where an issue or multiple issues occurred. A compressor or pump specialist with the vendor then reviews the output from the predictive model and provides back a report to the operator that summarizes what happened with the equipment and any probable issues that were found.
This allows an operator to evaluate the outcome without going all in and gives them the option to proceed in a more phased approach introducing predictive analytics to their asset performance management strategy and scaling through the organization over time.
Jeremy is the Director, Operations Marketing Portfolio which includes AVEVA’s HMI, SCADA, Enterprise Visualization and Value Chain Optimization portfolio. He has 18 years experience working with software and hardware companies spanning marketing, sales, product management and development roles focused in oil & gas, telecom and high performance computing. Jeremy enjoys taking complex concepts and breaking them down into understandable insights and stories. Jeremy has a Bachelor and Master of Science in Electrical Engineering with the University of Calgary and an MBA with the Smith School of Business at Queen's University.